The Science of Product Development: Bringing Causal Inference to Conversion and Retention Metrics by David Robinson
Visit rstats.ai for information on upcoming conferences.
Abstract: Modern websites track every pageview and click that their users perform, and have a strong interest in using that data to discover friction and smooth the journey. So then why are so many websites still so hard to use? I’ll make the case that the problem is largely a scientific one: even when we have the right data, we lack the conceptual and statistical tools to draw causal conclusions about user behavior. In this talk, I’ll lay out an early vision of what “product science” could be. I’ll introduce journeygrams, a method for quantifying and reasoning about sequential user behavior, and show how they can make product concepts like friction, backtracking, and retention more rigorous and actionable. I’ll include some examples of how typical product problems should be analyzed, and why our new approach is better suited to these problems than classical statistics and ML. These principles could help anyone looking to use data to improve their own products, and I hope will contribute to bringing the causal revolution to product development.
Bio: David Robinson is Director of Data Science at Contentsquare, where he's helping to build the next generation of product analytics technology. He's the co-author with Julia Silge of the tidytext package and the O’Reilly book Text Mining with R. He also created the broom, fuzzyjoin, and widyr packages, and authored the e-book Introduction to Empirical Bayes. David is passionate about R, statistics, education, cocktails, products, Taylor Swift, probability, and his two children.
Twitter: / drob
Presented at the 2024 New York R Conference (May 16, 2024)
Hosted by Lander Analytics (landeranalytics.com)
Негізгі бет Ғылым және технология David Robinson - Science of Prod Devel: Bringing Causal Inference to Conversion & Retention Metrics
Пікірлер: 1